Non-thermal plasma driven dry reforming of methane: electron energy-input power coupling mechanism and catalyst design criteria

Abstract

Dielectric barrier discharge plasma-driven dry reforming of methane is a promising technology for syngas production. However, plasma involves complex chemical reaction pathways, non-thermal equilibrium kinetic characteristics, and interactions with catalysts, which together affect the catalytic efficiency of the dielectric-barrier plasma driven dry reforming of methane reaction and constitute its main technical challenges. This study systematically investigates the effect of critical parameters-including reactor dimensions, input power, gas flow rate, gas composition, and catalyst type-on CH₄ and CO₂ conversion as well as syngas selectivity. Through thermodynamic and kinetic analysis, we elucidate the stepwise evolution mechanism of CH₄/CO₂ reactions under low-temperature plasma conditions. Notably, we incorporated the power law relationship between electron energy and input power into the thermodynamic model, thereby quantitatively revealing for the first time the regulatory effect of input power on the reaction path. This study provides valuable design principles to enhance the efficiency and industrial applicability of dielectric-barrier plasma driven dry reforming of methane processes.

Graphical abstract

Keywords

dry reforming of methane / plasma / non-thermal kinetics / nickel catalyst / thermodynamic

1 Introduction

Dry reforming of methane (DRM), an efficient hydrogen production method, is widely used in hydrogen production and gas synthesis. Spurred by the growing demand for energy transformation and environmental protection, methane dry reforming not only helps to optimize the utilization efficiency of natural gas resources but also provides a feasible way for the development of a low-carbon economy [1,2]. However, the traditional DRM process suffers from drawbacks such as high energy consumption, high temperature, and dependence on catalysts, consequently hampering wider industrial adoption [3−5].

Current industrial DRM technology mainly relies on high-temperature gasification, which not only has high energy consumption but also requires equipment with resistance to high temperature [6,7]. Therefore, reducing energy consumption and improving reaction efficiency are crucial to the future development of DRM. In this respect, dielectric barrier discharge (DBD) plasma-driven DRM is an emerging plasma-based catalytic reaction technology because it can efficiently excite CH₄ and CO₂ at room temperature and pressure. By using suitable catalysts, the reaction selectivity and conversion rate can be improved to potentially overcome the aforementioned hurdles [8,9].

Although DBD-DRM has been demonstrated in the laboratory, the low contact efficiency between the plasma and catalyst, low energy utilization efficiency, and complex product composition hinder further development. Simulation studies [10] have shown that in the normal pressure DBD reaction process, electrons will split CO₂ through electron impact dissociation and produce hundreds of reactions. At present, Ni-based catalysts are one of the common catalysts in plasma DRM processes [11,12]. Although methane conversion and hydrogen production have been improved, the energy consumption and catalytic efficiency are still inferior [13,14]. The energy efficiency (E_eff) of CO₂ decomposition by DBD is usually less than 15% primarily because the dissociation process is dominated by electron-impact reactions, which are inherently energy-inefficient. In these non-thermal plasmas, only a small fraction of the input electrical energy is transferred to high-energy electrons capable of dissociating CO₂ molecules. Moreover, a significant portion of the energy is lost through elastic collisions, excitation of vibrational and rotational states, and other non-dissociative pathways, all of which contribute to the overall low energy utilization efficiency [15−17]. In addition, the DBD reactor design, optimization of operating conditions, and synergistic effects of plasma and catalytic reactions require improvement and better understanding to address industrial needs [18−20]. The input power directly regulates the discharge mode and the energy distribution of electrons in the plasma, thereby affecting the types and quantities of reactive species, and is the core parameter that affects the dissociation path and reaction rate of CO₂ and CH₄. The gas composition (especially the CH₄/CO₂ molar ratio) determines the relative concentration and stoichiometric balance of the reactants, and has an important influence on the composition of the syngas and the tendency of side reactions (such as carbon deposition). The gas flow rate affects the residence time of the reactants in the discharge zone, thereby affecting the degree of reaction and energy utilization efficiency. The reactor structure, especially the choice of electrode spacing, electrode material, and insulating medium, will change the discharge uniformity and plasma volume, thereby affecting the stability of the plasma and the spatial distribution of active species. In terms of catalysts, their metal components, carrier properties and metal-carrier interactions determine the surface reactivity and carbon resistance, especially in the synergistic effect of plasma and catalyst. Therefore, the optimization of the DBD-DRM system needs to comprehensively consider the above factors to achieve higher reaction selectivity, E_eff and catalytic stability. Among these factors, input power plays a critical role in determining the dominant reaction pathways by influencing electron energy distribution, plasma density, and the generation of reactive species. Therefore, a systematic and quantitative investigation of input power, alongside other key parameters such as reactor size, gas flow rate, gas ratio, and catalyst selection, is essential for gaining deeper insights into the reaction mechanisms and for guiding the rational design of energy-efficient DBD-DRM systems.

In the DBD-DRM process, the reaction path involves multiple endothermic and exothermic steps [21,22]. The main endothermic reaction includes the direct dissociation of CH₄ and CO₂. The exothermic step mainly involves the formation of stable products by free radical recombination. Since the DBD reaction is non-thermal plasma driven, the activation of the reactants does not depend on the overall temperature, but is driven by high-energy electrons to excite the molecules into a highly reactive state [23]. Non-thermal kinetic models are crucial in this process and can be used to describe the electron collision cross section under plasma action, electric field-driven free radical generation, reaction rate, and product distribution, providing theoretical support for optimizing the specific energy input (SEI), gas flow ratio, and catalyst synergy [24,25]. Although studies have clarified the coexistence of endothermic/exothermic reactions in DBD-DRM [21,22,26], the regulation mechanism of how to quantitatively correlate the electron energy distribution and SEI and how to inhibit and improve the efficiency of syngas production through catalyst structure design are still technical bottlenecks. Traditional thermodynamic models are difficult to describe the nonlinear relationship between electron energy and reaction rate in non-equilibrium states in DBD, and the matching mechanism between catalyst pores and plasma penetration has not yet been clarified.

This work studies the effects of various reaction parameters on the conversion rate and selectivity of DBD methane dry reforming reaction, evaluates the technical feasibility of DBD-DRM, and explores the optimization path. Based on non-equilibrium thermodynamics and kinetics analysis, the gradual evolution mechanism of the DBD-DRM reaction is discussed, and a non-thermal kinetic rate model is established to quantify the regulatory effect of input power on the reaction path. In addition, combining experimental data and model analysis, a critical pore size criterion is proposed to provide theoretical guidance for the design of efficient catalysts.

2 Experimental

2.1 Catalyst synthesis

The catalysts were prepared by the impregnation method using equal volumes of solvent. Ni-loaded γ-Al₂O₃ and Ni-loaded SiO₂ were compared to evaluate the effects of different supports. The substrate (γ-Al₂O₃/SiO₂, 100 mg, purity 99.99%) was dispersed in 25 mL of deionized water by sonication for 30 min. To load 20 wt % nickel, 0.068 mg of Ni(NO₃)₂·6H₂O (AR, purity 98%) was dissolved in 25 mL of deionized water, stirred until fully dissolved, and then gradually added to the support solution with a 1:1 volume ratio. The mixture was stirred at room temperature for 4 h, followed by 12 h of standing to ensure complete nickel adsorption. The catalyst was washed with deionized water three times by centrifugation until pH 7 and then dried at 90 °C under vacuum for 12 h. The dried catalyst was calcined in a muffle furnace at 500 °C for 2 h with a heating rate of 1.5 °C·min^–1 to decompose the nitrate. Prior to use, the catalyst was reduced in a 20 mL·min^–1 H₂ atmosphere at a heating rate of 2 °C·min^–1 to 700 °C for 2 h to ensure complete reduction of nickel oxide to nickel.

2.2 Catalyst characterization

The X-ray diffraction (XRD) patterns were collected using a BRUKER D8 ADVANCE with Cu Kα radiation (λ = 1.540598 Å), operating at 40 kV and 40 mA, with a scanning rate of 2°·min^–1 over a 5°–90° range. Surface morphology and particle size were characterized by scanning electron microscopy (SEM) at an accelerating voltage of 10–20 kV. Mapping mode was used for element analysis, with results normalized and repeated four times. Raman scattering was conducted in situ using a Horiba Scientific-LabRAM HR Evolution system. UV-Vis/NIR absorption spectra were obtained with a Hitachi UH4100 spectrometer, scanning in the 200–2000 nm range. X-ray photoelectron spectroscopy (XPS) was performed using a Thermo Scientific ESCALAB 250Xi with Al Kα radiation (1487.20 eV) and calibration at 284.6 eV for the C 1s peak. The temperature distribution during the reaction was recorded using a FLIR E6xt infrared thermal imager. H₂ temperature-programmed reduction (H₂-TPR) and CO₂ temperature-programmed desorption (CO₂-TPD) experiments were carried out using a Micromeritics AutoChem 2920 chemisorption analyzer. For each measurement, approximately 100 mg of the catalyst sample was pretreated under an argon flow (40 mL·min^–1) at 200 °C for 1 h, followed by cooling to room temperature (25 °C). For the H₂-TPR analysis, the sample was then exposed to a 10% H₂/Ar gas mixture (40  mL·min^–1), and the temperature was ramped from 25 to 800 °C at a heating rate of 10 °C·min^–1 to record the H₂ consumption profile. Similarly, for the CO₂-TPD analysis, the pretreated catalyst was cooled to 25 °C and saturated with high-purity CO₂ for 1 h. After saturation, the system was purged with argon (30 mL·min^–1) to remove physisorbed CO₂, followed by a temperature-programmed desorption process from 25 to 700 °C at a ramping rate of 10 °C·min^–1 to obtain the CO₂-TPD profile.

2.3 Experimental setup

The reduced catalyst (100 mg) was mixed with γ-Al₂O₃ or quartz sand in a 1:9 ratio and ground uniformly. It was then evenly dispersed in the center of the DBD reactor, with quartz wool used to fix the catalyst layer, forming a sandwich structure. CH₄ and CO₂ were introduced as reaction gases, with He serving as the carrier gas. After 10 min of gas flow, the plasma generator (CTP-2000K, Coronalab, China) was activated and set to the specified voltage and current. After stabilization, product analysis was carried out. After dehydration using silica gel, the products were analyzed by online gas chromatography (6890A, Agilent Technologies, USA) with 3–4 measurements per sampling group. The outlet flow was measured with a soap film flow meter, and the temperature distribution during the reaction was monitored with the infrared thermal imager.

A cylindrical DBD reactor system was built, with quartz glass as the dielectric barrier material (Fig. S1, cf. Electronic Supplementary Material, ESM). Three reactor sizes were designed to study the effect of reactor dimensions on DBD discharge: small (7 mm × 2 mm × 100 mm, discharge gap 3 mm), medium (16 mm × 2.5 mm × 150 mm, discharge gap 4 mm), and large (20 mm × 2.5 mm × 150 mm, discharge gap 2 mm). For energy consumption and performance comparison, a constant input current of 1 A was used, with a discharge frequency of 7–10 kHz and input voltage ranging from 30 to 100 V, corresponding to input power levels between 30 and 100 kW. The high-voltage electrode was a stainless-steel tube at the reactor’s center, with a stainless steel mesh fixed outside the quartz tube to act as the ground electrode for discharge dispersion. The discharge voltage and current signals were recorded using a Rigol DS2302A oscilloscope (China).

2.4 DBD catalysis evaluation

To evaluate the performance of DBD-DRM, the SEI (kJ·L^–1) was calculated by the following formula to reflect the effects of discharge power and total flow rate:

(1)SEI=\frac{60\times W_{\mathrm{I}\mathrm{n}\mathrm{t}}}{Q},

where W_Int is the discharge input power (W), and Q is the total flow rate (mL·min^–1). The conversion rate of CO₂ and CH₄ is calculated using the following formula:

(2)C_{{\mathrm{C}\mathrm{O}}_2}=\frac{\left(C_{{\mathrm{C}\mathrm{O}}_2}-C_{{\mathrm{C}\mathrm{O}}_2,\mathrm{o}\mathrm{u}\mathrm{t}}\right)}{C_{{\mathrm{C}\mathrm{O}}_2}}\times 100,

(3)C_{{\mathrm{C}\mathrm{H}}_4}=\frac{(C_{{\mathrm{C}\mathrm{H}}_4}-C_{{\mathrm{C}\mathrm{H}}_4,\mathrm{o}\mathrm{u}\mathrm{t}})}{C_{{\mathrm{C}\mathrm{H}}_4}}\times 100,

where C_{{\mathrm{C}\mathrm{O}}_2} and C_{{\mathrm{C}\mathrm{H}}_4} are the CO₂ and CH₄ concentrations (mol·s^–1) of the inlet gas, respectively, C_{{\mathrm{C}\mathrm{O}}_2,\mathrm{o}\mathrm{u}\mathrm{t}} and C_{{\mathrm{C}\mathrm{H}}_4,\mathrm{o}\mathrm{u}\mathrm{t}} are the CO₂ and CH₄ concentrations (mol·s^–1) measured by GC, respectively. The yield (Y) and selectivity (S) of the gas product are determined by the following method:

(4)Y_{{\mathrm{H}}_2}=\frac{C_{{\mathrm{H}}_2}}{2\times C_{{\mathrm{C}\mathrm{H}}_4}}\times 100,

(5)Y_{\mathrm{C}\mathrm{O}}=\frac{C_{\mathrm{C}\mathrm{O}}}{C_{{\mathrm{C}\mathrm{O}}_2}+C_{{\mathrm{C}\mathrm{H}}_4}}\times 100,

(6)S_{{\mathrm{H}}_2}=\frac{C_{{\mathrm{H}}_2}}{2\times (C_{{\mathrm{C}\mathrm{H}}_4}-C_{{\mathrm{C}\mathrm{H}}_4,\mathrm{o}\mathrm{u}\mathrm{t}})}\times 100,

(7)S_{\mathrm{C}\mathrm{O}}=\frac{C_{\mathrm{C}\mathrm{O}}}{\left(C_{{\mathrm{C}\mathrm{H}}_4}-C_{{\mathrm{C}\mathrm{H}}_4,\mathrm{o}\mathrm{u}\mathrm{t}}\right)+\left(C_{{\mathrm{C}\mathrm{O}}_2}-C_{{\mathrm{C}\mathrm{O}}_2,\mathrm{o}\mathrm{u}\mathrm{t}}\right)}\times 100.

The carbon balance (B) of the DBD-DRM reaction process is calculated as follows:

(8)B_{{\mathrm{C}}_x{\mathrm{H}}_y}=\frac{x\times C_{{\mathrm{C}}_x{\mathrm{H}}_y}}{\left(C_{{\mathrm{C}\mathrm{H}}_4}-C_{{\mathrm{C}\mathrm{H}}_4,\mathrm{o}\mathrm{u}\mathrm{t}}\right)+\left(C_{{\mathrm{C}\mathrm{O}}_2}-C_{{\mathrm{C}\mathrm{O}}_2,\mathrm{o}\mathrm{u}\mathrm{t}}\right)}\times 100.

Among them, C_{{\mathrm{H}}_2}, C_{\mathrm{C}\mathrm{O}} and C_{{\mathrm{C}}_x{\mathrm{H}}_y} are the concentrations of H₂, CO, and C_{{\mathrm{C}}_x{\mathrm{H}}_y} measured by GC (mol·s^–1).

2.5 Plasma discharge calculation

Under the action of the alternating electric field, it is difficult to measure the discharge power because the phase difference between voltage and current is difficult to determine. Therefore, the voltage-charge Lissajous figure area integration method is used to determine the power of the dielectric barrier discharge. The maximum RF power and stable Lissajous figure of the oscilloscope are obtained by adjusting the discharge frequency. The input power and plasma conversion rate will also fluctuate. The actual discharge power P_act of DBD is calculated as follows:

(9)P_{\mathrm{a}\mathrm{c}\mathrm{t}}=\frac{1}{T}{\int }_0^{T}VI\mathrm{d}t=\frac{C_M}{T}{\int }_0^{T}V\frac{{\mathrm{d}V}_M}{\mathrm{d}t}=f{C}_M\int {V\mathrm{d}V}_M=fS.

In this formula: P is the discharge power (W); T is the period of the AC power supply (s); I is the loop current (A); f is the frequency of the applied voltage (Hz). S is obtained by integrating the area of the Lissajous figure.

For the E_eff evaluation of DBD methane dry reforming, the following calculation formula is used:

(10)E_{\mathrm{C}}=\frac{P_{\mathrm{a}\mathrm{c}\mathrm{t}}}{W_{\mathrm{I}\mathrm{n}\mathrm{t}}}\times 100\mathrm{\%},

(11)E_{\mathrm{e}\mathrm{f}\mathrm{f}}=\frac{C_{{\mathrm{C}\mathrm{O}}_2,\mathrm{c}\mathrm{o}\mathrm{n}}+C_{{\mathrm{C}\mathrm{H}}_4,\mathrm{c}\mathrm{o}\mathrm{n}}}{P_{\mathrm{a}\mathrm{c}\mathrm{t}}}\times 100\mathrm{\%}.

Among them, E_C is the energy conversion efficiency of the system, and E_eff is the energy yield. A measuring capacitor (C_M) is connected in series between the grounding electrode of the reactor and the ground, and the voltage across the reactor (V) and the voltage across the measuring capacitor (V_M) are collected. The two channels of the oscilloscope are connected and adjusted to the X-Y mode to obtain a closed curve, which is the Lissajous figure, as shown in Fig. S2 (cf. ESM).

3 Results and discussion

3.1 Parameters affecting the DBD reactor size

The reactor dimensions (e.g., discharge gap, diameter, and length) affect the plasma discharge characteristics and electric field distribution, consequently impacting the efficiency of generating active species, reaction rates, and product selectivity. By analyzing the effects of the reactor size, energy utilization, and methane conversion can be optimized for energy-efficient, low-consumption industrial-scale DBD reactors. Here, the performance of three DBD reactors with varying sizes is compared in Fig.1(a)–Fig.1(c), The detailed parameters are shown in Fig.1(d). To estimate the effective discharge volume (V_eff) of plasma generation in three different sizes of DBD reactors, this study adopts an approximate calculation method based on geometric parameters and discharge filling factor (f). Specifically, it is assumed that plasma is effectively generated only in part of the discharge gap, the discharge filling factor is 0.3, and the formula V_eff = f × A × d is used for calculation in combination with the electrode area and spacing. The results show that the effective discharge volume of the small-sized reactor is V_eff = 6.37 × 10⁻⁹ m³, the medium-sized reactor is 1.14 × 10⁻⁷ m³, and the large-sized reactor is 1.06 × 10⁻⁷ m³. These results can provide a reference basis for the analysis of energy density and plasma reaction behavior of reactors of different sizes.

As shown in Fig.1(e), methane conversion increases from 30.96% for the small reactor to 36.29% for the large reactor. In contrast, CO₂ conversion rises slightly from 8.71% to 9.66%. This minor increase is attributed to the smaller discharge gap and optimal dielectric barrier thickness in the larger reactor, leading to more uniform and stable discharge. This improves the reaction activity and methane conversion efficiency. However, CO₂ conversion is limited by chemical equilibrium and side reactions, resulting in a smaller increase. The yields of H₂ and CO also increase significantly with reactor size. The H₂ yield rises from 9.27 to 12.54, while the CO yield increases from 4.6 to 11.32, indicating that a larger reactor provides a more stable discharge environment to facilitate methane cracking and CO₂ reactions. Higher discharge power and stable discharge characteristics also enhance the probability of reactants participating in the main reactions and reduce the occurrence of inefficient side reactions. Furthermore, the H₂ and CO selectivity increases with reactor size. The H₂ selectivity increases from 14.97% to 18.43%, while the CO selectivity rises from 6.17% to 14.06%. Meanwhile, the selectivity of by-products, such as C₂H₄, C₂H₆, and C₃H₈, decreases, particularly C₂H₄ which shows a decrease from 0.44% to 0.23%. The selectivity of C₂H₆ is very low (less than 4%), and its yield is correspondingly minor, confirming that higher hydrocarbons are not dominant products in this system. The results suggest that DBD reactors with larger effective discharge volumes and relatively smaller discharge gaps facilitate more homogeneous and volume-dominated plasma generation, which enhances the uniformity of the energy distribution within the reaction zone. The enlarged plasma-active region allows for a higher density of low-to-medium energy electrons, which are more favorable for the selective recombination of radicals into small molecular products such as CH₄ and CO₂, while minimizing excessive fragmentation and polymerization reactions that lead to the formation of complex hydrocarbons. Moreover, the reduced discharge gap strengthens the electric field, thereby promoting efficient excitation and dissociation of gas molecules at lower power inputs, contributing to improved reaction control and higher selectivity toward desired products.

Power measurements show that the small reactor operated at 14.9 and 16.5 W and the large reactor at 19.9 W, demonstrating higher discharge efficiency in the larger reactor for the same input power. This efficiency stems from that the smaller discharge gap and optimal dielectric barrier thickness contribute to more uniform and stable discharges for better efficiency. Additionally, the higher discharge power increases the average energy of electrons, thus enhancing the concentration of reactive species (such as electrons, ions, and free radicals) and promoting the main reactions. The decrease in byproduct selectivity indicates that the reaction pathway is optimized, reducing the formation of complex hydrocarbons. However, the increase in size also brings about an increase in energy requirements and an increase in design complexity.

3.2 Effects of gas flow ratio and energy input on CO₂ and CH₄ conversion efficiency

The conversion rates of CO₂ and CH₄ vary significantly for different gas flow ratios. When only CO₂ is present, the conversion rate is 17.78%, while CH₄ shows a conversion rate of 51.05%. This highlights the limited reaction efficiency of single gases, particularly CO₂, which shows lower conversion in the absence of reducing gases like CH₄. However, when CO₂ and CH₄ are introduced at flow rates of 10 mL·min^–1 each (i.e., a CH₄:CO₂ ratio of 1:1 with a total flow rate of 20 mL·min^–1), the CH₄ conversion reaches 55.37%. Increasing the flow rates to 15 mL·min^–1 each (total 30 mL·min^–1) further enhances the CH₄ conversion to 68.95%. However, when the flow rates increase to 20 mL·min^–1 each (total 40 mL·min^–1), the conversion rates of CO₂ and CH₄ drop to 14.16% and 50.10%, respectively. This decline is likely due to overly high reactant concentrations, which lead to uneven energy distribution in the plasma zone, hinder the formation of reactive species, and promote side reactions (Fig.2). The plasma input power in all cases was maintained at 30 W.

The introduction of He as a carrier gas alters the conversion rates of CO₂ and CH₄. Under the condition of CO₂:CH₄:He = 20:20:20, the conversion rates of CO₂ and CH₄ are 18.20% and 51.16% and are slightly higher than those without the carrier gas. When the carrier gas ratio is further increased (e.g., CO₂:CH₄:He = 15:15:30 or 6:6:48), the conversion rates increase significantly, with the highest conversion (41.89% and 66.75%, respectively) observed for CO₂:CH₄:He = 6:6:48. The observation suggests that He distributes the plasma energy more evenly, enhances the discharge stability, and reduce side reactions, especially under high-dilution conditions [27,28].

The optimal CO₂/CH₄ ratio maximizes the synergistic effect, and He, as a carrier gas, optimizes the reaction conditions. However, high reactant concentrations may cause energy distribution imbalances or increased side reactions to lower the conversion efficiency. The dilution effect of the carrier gas improves plasma discharge and enhances active species generation. While He can enhance discharge uniformity and reactivity in DBD systems, its practical application faces significant drawbacks. Industrially, He separation from product streams requires energy-intensive processes like membrane separation or cryogenic techniques, increasing operational costs. Additionally, its high proportion is needed to notably improve CO₂/CH₄ conversion, leading to substantial waste and economic inefficiency. Analytically, He interferes with H₂ detection in gas chromatography due to peak overlap, complicating yield measurements. Alternative methods—such as coupling thermal conductivity detectors with mass spectrometry or using Ar as a carrier gas—may address this, but risk higher instrument costs or reduced sensitivity. Thus, despite its plasma-enhancing properties, He’s high cost, separation challenges, and analytical limitations necessitate a careful balance between performance gains and economic viability.

In large DBD reactors, the conversion rates of CH₄ and CO₂ are influenced by the SEI and CH₄/CO₂ feed ratio (Fig.3(a, b)). As SEI increases from 90 to 240 J·mL^–1, both the CH₄ and CO₂ conversion rates rise. A higher SEI provides greater energy density, generates more active species, and enhances reaction efficiency. For a 1:1 CH₄/CO₂ feed ratio, CH₄ conversion increases from 34.66% to 61.26%, while CO₂ conversion increases from 9.05% to 21.48%. This change indicates that higher energy inputs promote the generation of energetic electrons and active species to improve reaction efficiency.

The CH₄/CO₂ feed ratio also significantly affects the conversion rates. Under high CH₄ concentrations (e.g., 4:1 and 9:1), CH₄ conversion is higher, while CO₂ conversion is lower (Fig.3(a)). Under low SEI conditions, CO₂ conversion is quite low, in fact, almost zero when the CH₄/CO₂ ratio is 1:4 or 1:9, because the limited active species generated at low energy levels make CO₂ decomposition more difficult. In contrast, at lower CH₄ concentrations, the electron energy in the DBD plasma is less consumed by CH₄ dissociation and excitation processes, which typically require relatively high electron energies. As a result, more electrons remain available to activate CO₂ molecules, enhancing their conversion efficiency. Additionally, the lower CH₄ content leads to a more balanced generation of reactive species such as O•, CO, and H• radicals, and reduces the spatial and temporal fluctuation in local electron density and temperature. This results in a more uniform plasma environment across the discharge region, where energy is more evenly distributed. The improved uniformity of active species generation minimizes localized over-reduction or excessive carbon formation, thereby favoring a more selective and efficient CO₂ activation pathway [20,29,30]. The nonlinear characteristics of SEI also manifest at different feed ratios. At 240 J·mL^–1, CH₄ conversion significantly increases, particularly for large CH₄/CO₂ ratios (e.g., 9:1), reaching 78.38%, while CO₂ conversion is highest for a 1:1 ratio, reaching 21.48%. Under lower SEI conditions (e.g., 90 J·mL^–1), CH₄ conversion increases slowly (Fig.3(b)), and CO₂ conversion remains low, indicating that low SEI limits active species generation and reduces reaction efficiency. As the energy input increases, the number of high-energy electrons and reactive species (such as O, OH, H, etc.) produced in the reactor increases, making CO₂ decomposition and related reactions (such as the reverse water gas shift reaction) more efficient.

The coupling effect of the CH₄/CO₂ ratio and SEI reflects the complex dynamic behavior. Under low SEI (90 J·mL^–1), high CH₄ ratios (e.g., 9:1) maintain relatively high conversion rates, suggesting that higher methane concentration promotes the collision frequency (Fig.3(a) and Fig.3(b)). However, under high SEI conditions (240 J·mL^–1), the conversion rate becomes more sensitive to feed ratio changes, indicating that a higher energy input enhances the responsiveness to feed ratios. Optimizing dry methane reforming requires balancing CH₄ and CO₂ conversion rates. Low CH₄/CO₂ ratios (e.g., 1:1 or 3:7) achieve reasonable conversion rates at a high SEI, but produce more side products, whereas high CH₄/CO₂ ratios (e.g., 9:1) increase CH₄ conversion but lower CO₂ utilization efficiency. In practical applications, adjusting SEI, optimizing the feed ratio, or introducing co-catalysts can improve the overall reaction efficiency.

The yield and selectivity of H₂ and CO exhibit clear trends under varying SEI and CH₄/CO₂ ratios. As shown in Fig.3(c), the H₂ yield generally increases with increasing SEI, indicating that higher energy input promotes CH₄ activation and hydrogen production. This effect is more pronounced under CH₄-rich conditions (e.g., 4:1 and 9:1), where sufficient CH₄ provides a greater source of hydrogen-containing fragments. The maximum H₂ yield is observed at SEI = 240 and CH₄/CO₂ = 9:1. In contrast, at low CH₄/CO₂ ratios (e.g., 1:9), the H₂ yield remains low across all SEI values, especially under low SEI, due to limited hydrogen source and insufficient energy to effectively drive CH₄ cracking. Therefore, rather than concluding that low power input is more favorable, these results demonstrate that a combination of high CH₄ concentration and adequate SEI is necessary for maximizing hydrogen production via methane decomposition pathways.

The CO yield also increases with increasing SEI, but the magnitude is not as significant as that of H₂, as shown in Fig.3(d). The generation of CO mainly depends on the participation of CO₂. At high CO₂ ratios (such as 1:9 and 1:4), the CO yield is high and reaches a peak in the system with CH₄/CO₂ of 7:3. On the contrary, in systems with high CH₄ content (such as 4:1 and 9:1), the CO yield is relatively low, especially when the SEI is low. In terms of selectivity, the selectivity of H₂ increases with the increase of SEI, and the improvement is more significant in systems with high CH₄ content (such as 9:1 and 4:1), as shown in Fig.3(e). This shows that the cracking of CH₄ plays a dominant role in the generation of H₂. When the CH₄/CO₂ ratio is 9:1, the H₂ selectivity can exceed 80%, while when the CH₄/CO₂ ratio is 1:9, the H₂ selectivity is the lowest, and even less than 10% when the SEI is low. The CO selectivity is closely related to the CO₂ ratio. In high CO₂ ratio systems (such as 1:9 and 1:4), the CO selectivity is high, especially close to 100% under high SEI conditions. In systems with high CH₄ content, the CO selectivity is significantly reduced, as shown in Fig.3(f).

The H₂ yield increases with SEI, peaking at SEI = 240 and CH₄/CO₂ = 9:1, particularly in systems with high CH₄ ratios (e.g., 4:1, 9:1). In CO₂-deficient systems (e.g., 1:9), H₂ yield is lowest, especially for a low SEI (Figs. S3(a, b), cf. ESM). The CO yield also rises with SEI but to a lesser extent, with higher CO₂ ratios (e.g., 1:9, 1:4) promoting CO generation. The H₂ selectivity increases with SEI, especially in high CH₄ ratio systems, reaching over 80%, while CO selectivity is higher at high CO₂ ratios, nearing 100% under high SEI. The CO₂/CH₄ ratio significantly influences conversion rates, selectivity, and CO/H₂ molar ratios. CO₂ conversion increases initially and then decreases with higher CO₂/CH₄ ratios, while CH₄ conversion remains high. The CO yield peaks at a CO₂/CH₄ ratio of 2.33, and the CO/H₂ molar ratio rises with the CO₂/CH₄ ratio. A lower power favors CH₄ cracking for H₂ production.

The reaction mechanisms reveal that CH₄ cracking is key to H₂ generation (CH₄ → C + 2H₂), while CO₂ reduction dominates CO production (CO₂ + CH₄ → 2CO + 2H₂). A higher SEI accelerates both reactions to enhance the H₂ and CO yields, whereas a low SEI limits these reactions, especially with extreme feed ratios (Fig. S3(c)). The CO₂/CH₄ ratios play a critical role in industrial processes. Fischer-Tropsch synthesis is best for ratios of 1.5–2, methanol synthesis and syngas at 1, and ammonia synthesis requires a CO/H₂ ratio below 0.1 (Fig. S3(d)). Therefore, adjusting the feed ratios, SEI, and CO₂/CH₄ ratios optimizes the process for diverse industrial applications.

3.3 Effects of input power and fillers on CO₂ and CH₄ conversion rates

Both the CO₂ and CH₄ conversion rates increase with higher input power, as a higher power provides more energy to ionize and excite reactant molecules [31,32], thus enhancing plasma chemical reactions (Fig.4(a)). In the absence of fillers, the conversion rates are relatively low due to the limited discharge region. When γ-Al₂O₃ is used as a filler, its larger surface area and excellent thermal stability promote the adsorption of reactant molecules and surface reactions, leading to a CO₂ conversion rate of 65.32%. Quartz sand, with good electrical insulation, enhances the plasma uniformity and improves both the CO₂ and CH₄ conversion rates [33]. Silica powder fares better at a low power but shows lower conversion at a high power, likely due to the adsorption of excessive intermediate products [34].

Different fillers affect the actual energy conversion rate (Fig.4(b)) and the E_eff. Without fillers, the actual energy conversion rate increases with power, but E_eff decreases as the power increases (Fig.4(c)). At high power, γ-Al₂O₃ shows a conversion rate of 69%, but E_eff gradually decreases with increasing power. Quartz sand maintains a stable energy conversion rate of 75.72% at 100 W, with a relatively stable E_eff, making it suitable for higher power applications. The electrical insulation properties and high thermal stability of quartz sand help enhance the uniformity and reaction efficiency of plasma, but its catalytic performance is relatively weak. Silica powder is better at lower power, but E_eff decreases at higher power, indicating reduced E_eff. Silica powder has a moderate specific surface area and can maintain high reactivity at low energy input. However, as the power increases, its E_eff gradually decreases from 0.54021 at 40 W to 0.26956 at 100 W, indicating that the marginal efficiency of energy utilization at high power is weakened.

Figure S4 (cf. ESM) shows infrared thermal images of the reactor with and without fillers at a CH₄:CO₂ feed ratio of 15:15 mL·min^–1. The temperature in the center of the reactor rises significantly with increasing power. After collisions between high-energy particles and gases, energy is dissipated as heat, leading to a reduction in conversion efficiency. When the high-frequency particles generated by the high-voltage electrode diffuse to the low-voltage electrode, the energy is transferred through the collision of high-concentration CO₂ and CH₄, and the energy will gradually decrease. Therefore, the e-near the center area will have lower energy and cannot further convert CO₂ and CH₄, and can only be converted in the form of heat. Due to the complex conversion process of CO₂ and CH₄ in the plasma electric field, a large number of neutral and charged particles will be generated, and this state is instantaneously stable in the electric field. However, during the mutual collision reaction, it cannot be maintained for a long time, and the reverse reaction returns to the steady-state of CO₂ and CH₄, and the remaining exothermic reactions will occur, which will be dissipated in the form of heat.

Overall, the input power and filler type influence the CO₂ and CH₄ conversion rates. Quartz sand and γ-Al₂O₃ are the best under high-power conditions, rendering them suitable for industrial applications, while silica powder is more effective under low-power conditions, making it suitable for low-energy scenarios.

3.4 Analysis of DBD-DRM reaction mechanism based on thermodynamics and kinetics

By analyzing the main endothermic and exothermic reactions of CO₂ and CH₄ decomposition in DBD [26], it was found that the plasma reaction follows a non-equilibrium mechanism and is not completely controlled by ∆H. Factors such as electron collisions and active free radicals may affect the reaction sequence. Under low electron energy (1–5.5 eV) conditions, low-energy electrons preferentially break C–H (CH₄) and C=O (CO₂) bonds through vibrational excitation (such as vibrational mode excitation of CO₂, ∆H = 0.083 eV) and direct electron collisions to generate free radicals (CH₃, H, O) and CO. The primary dissociation reactions (En1–En4) of CO₂ and CH₄ are mainly triggered by electron collisions. CO₂ dissociates through En1 and En4 to form CO and O species, while CH₄ mainly undergoes En2 and En3 reactions to break C–H bonds and generate CH₃ and H free radicals. The reaction rate at this stage is mainly controlled by the electron density (n_e) and the collision cross section (σ), and the vibration excitation mechanism plays a key role in the initial breakage of the C–H and C=O bonds. As the electron energy increases (> 12 eV), CH₄ and CO₂ are ionized (En5–En6), further forming CH₄⁺ and CO₂⁺. When the electron energy reaches 14–19 eV, dissociative ionization (En7–En8) occurs, producing more active charged fragments (CH₃⁺, O⁺, etc.), providing precursor species for subsequent exothermic recombination reactions. The following are the main endothermic and exothermic reactions of CO₂ and CH₄ in the DBD reaction (Fig.5(a)):

(1) endothermic reactions

En1: CO₂ decomposition

{\mathrm{C}\mathrm{O}}_2\to \mathrm{C}\mathrm{O}+1/2{\mathrm{O}}_2(\mathrm{\Delta }H=2.93\mathrm{e}\mathrm{V})

En2: CH₄ electron impact dissociation

{\mathrm{e}}^{-}+{\mathrm{C}\mathrm{H}}_4\to {\mathrm{e}}^{-}+{\mathrm{C}\mathrm{H}}_3+\mathrm{H}(\mathrm{\Delta }H=4.5\mathrm{e}\mathrm{V})

En3: CH₄ decomposition

{\mathrm{C}\mathrm{H}}_4\to {\mathrm{C}\mathrm{H}}_3+\mathrm{H}(\mathrm{\Delta }H=4.55\mathrm{e}\mathrm{V})

En4: CO₂ electron impact dissociation

{\mathrm{e}}^{-}+{\mathrm{C}\mathrm{O}}_2\to {\mathrm{e}}^{-}+\mathrm{C}\mathrm{O}+\mathrm{O}(\mathrm{\Delta }H=5.5\mathrm{e}\mathrm{V})

En5: CH₄ electron impact ionization

{\mathrm{e}}^{-}+{\mathrm{C}\mathrm{H}}_4\to {\mathrm{e}}^{-}+{\mathrm{C}\mathrm{H}}_4^{+}+{\mathrm{e}}^{-}(\mathrm{\Delta }H=12.6\mathrm{e}\mathrm{V})

En6: CO₂ electron impact ionization

{\mathrm{e}}^{-}+{\mathrm{C}\mathrm{O}}_2\to {\mathrm{e}}^{-}+{\mathrm{C}\mathrm{O}}_2^{+}+{\mathrm{e}}^{-}(\mathrm{\Delta }H=13.8\mathrm{e}\mathrm{V})

En7: CH₄ electron impact dissociation ionization

{\mathrm{e}}^{-}+{\mathrm{C}\mathrm{H}}_4\to {\mathrm{e}}^{-}+{\mathrm{C}\mathrm{H}}_3^{+}+\mathrm{H}+{\mathrm{e}}^{-}(\mathrm{\Delta }H=14.3\mathrm{e}\mathrm{V})

En8: CO₂ electron impact dissociation ionization

{\mathrm{e}}^{-}+{\mathrm{C}\mathrm{O}}_2\to {\mathrm{e}}^{-}+\mathrm{C}\mathrm{O}+{\mathrm{O}}^{+}+{\mathrm{e}}^{-}(\mathrm{\Delta }H=19.2\mathrm{e}\mathrm{V})

When enough free radicals and ions accumulate, exothermic reactions (Ex1–Ex8) begin to dominate the chemical evolution of the system. Low-energy electrons (< 10 eV) promote Ex1–Ex3, that is, {\mathrm{C}\mathrm{O}}_2^{+}, {\mathrm{C}\mathrm{H}}_4^{+}, and CO⁺ recombine with electrons to generate CO, O, CH₃, and H, reducing the charge density of the plasma system. In the Ex4–Ex8 reaction stage, free radical recombination becomes the dominant process, involving species such as CH₃, O, and H. H and O undergo recombination reactions (Ex6–Ex7) to form H₂ and O₂, while CO reacts with O (Ex8) to generate CO₂, which ultimately determines the ratio of the synthesis gas (H₂/CO) product. The competition between H₂ generation (Ex6) and CO₂ reduction (Ex8) affects the selectivity of the final synthesis gas, and appropriate adjustment of the electron energy distribution helps to optimize the product ratio.

(2) Exothermic reactions

Ex1: {\mathrm{C}\mathrm{O}}_2^{+} reacts with e^– to produce CO and O

{\mathrm{C}\mathrm{O}}_2^{+}+{\mathrm{e}}^{-}\to \mathrm{C}\mathrm{O}+\mathrm{O}(\mathrm{\Delta }H=-13.8\mathrm{e}\mathrm{V})

Ex2: {\mathrm{C}\mathrm{H}}_4^{+} reacts with e^– to form CH₃ and H

{\mathrm{C}\mathrm{H}}_4^{+}+{\mathrm{e}}^{-}\to {\mathrm{C}\mathrm{H}}_3+\mathrm{H}(\mathrm{\Delta }H=-12.6\mathrm{e}\mathrm{V})

Ex3: CO⁺ reacts with e^– to produce CO

{\mathrm{C}\mathrm{O}}^{+}+{\mathrm{e}}^{-}\to \mathrm{C}\mathrm{O}(\mathrm{\Delta }H=-14\mathrm{e}\mathrm{V})

Ex4: CH₃ reacts with CH₃ to generate C₂H₆

{\mathrm{C}\mathrm{H}}_3+{\mathrm{C}\mathrm{H}}_3\to {\mathrm{C}}_2{\mathrm{H}}_6(\mathrm{\Delta }H=-0.91\mathrm{e}\mathrm{V})

Ex5: O reacts with O₂ to generate O₃

\mathrm{O}+{\mathrm{O}}_2+\mathrm{M}\to {\mathrm{O}}_3+\mathrm{M}(\mathrm{\Delta }H=-1.09\mathrm{e}\mathrm{V})

Ex6: H reacts with H to produce H₂

\mathrm{H}+\mathrm{H}+\mathrm{M}\to {\mathrm{H}}_2+\mathrm{M}(\mathrm{\Delta }H=-4.52\mathrm{e}\mathrm{V})

Ex7: O reacts with O to generate O₂

\mathrm{O}+\mathrm{O}+\mathrm{M}\to {\mathrm{O}}_2+\mathrm{M}(\mathrm{\Delta }H=-5.16\mathrm{e}\mathrm{V})

Ex8: CO reacts with O to generate CO₂

\mathrm{C}\mathrm{O}+\mathrm{O}\to {\mathrm{C}\mathrm{O}}_2(\mathrm{\Delta }H=-5.51\mathrm{e}\mathrm{V})

According to different energy inputs, considering the main triggering energy (∆H) of CO₂ and CH₄ in the DBD reaction and their possible occurrence order in the plasma environment (Fig.5(b)), the reaction mechanism is as follows:

(3) priority triggering of primary endothermic reaction

In a low-energy electron environment (1–5.5 eV), En1–En4 occurs preferentially:

En2: CH₄ dissociation

{\mathrm{e}}^{-}+{\mathrm{C}\mathrm{H}}_4\to {\mathrm{e}}^{-}+{\mathrm{C}\mathrm{H}}_3+\mathrm{H}(k_1\propto {\sigma }_{{\mathrm{C}\mathrm{H}}_4}{n}_{\mathrm{e}}\sqrt{E_{\mathrm{e}}})

En4: CO₂ dissociation

{\mathrm{e}}^{-}+{\mathrm{C}\mathrm{O}}_2\to {\mathrm{e}}^{-}+\mathrm{C}\mathrm{O}+\mathrm{O}(k_2\propto {\sigma }_{{\mathrm{C}\mathrm{O}}_2}n{}_{\mathrm{e}}\sqrt{E{}_{\mathrm{e}}-5.5})

Low-energy electrons break C–H (CH₄) and C=O (CO₂) bonds through vibration excitation and electron collision, generating free radicals (CH₃, H, O) and CO.

(4) Medium-high energy ionization and dissociative ionization

When the electron energy is > 12 eV, En5–En8 triggers:

En5/En6: {\mathrm{C}\mathrm{H}}_4^{+}/{\mathrm{C}\mathrm{O}}_2^{+} ionization

k_{\mathrm{i}\mathrm{o}\mathrm{n}}=A\mathrm{e}\mathrm{x}\mathrm{p}\left(-\frac{\mathrm{\Delta }{H}_{\mathrm{i}\mathrm{o}\mathrm{n}}}{k_{B}T{}_{\mathrm{e}}}\right)(T_{\mathrm{e}}\gg T_{\mathrm{g}\mathrm{a}\mathrm{s}})

En7/En8: dissociative ionization (ΔH = 14–19 eV)

Need to be triggered by high-energy electrons (> 15 eV) to generate charged fragments ({\mathrm{C}\mathrm{H}}_3^{+}, O⁺, etc.).

(5) Dynamic equilibrium of exothermic reactions

Exothermic reactions dominate product distribution through ion recombination and free radical recombination. After the formation of active species, exothermic reactions dominate.

Ex1–Ex3: ion-electron recombination

{\mathrm{C}\mathrm{O}}_2^{+}+{\mathrm{e}}^{-}\to \mathrm{C}\mathrm{O}+\mathrm{O}(k_{\mathrm{r}\mathrm{e}\mathrm{c}}\propto n{}_{\mathrm{e}}{n}_{\mathrm{i}\mathrm{o}\mathrm{n}}{T}_{\mathrm{e}}^{-0.5})

Ex4–Ex8: free radical recombination

{\mathrm{C}\mathrm{H}}_3+{\mathrm{C}\mathrm{H}}_3\to {\mathrm{C}}_2{\mathrm{H}}_6(k_{\mathrm{r}\mathrm{a}\mathrm{d}}\propto T_{\mathrm{g}}^{-1}\mathrm{e}\mathrm{x}\mathrm{p}(-E_{\mathrm{a}}/k_B{T}_{\mathrm{g}}\left)\right)

The rates of H₂ generation (Ex6) and CO₂ regeneration (Ex8) determine the syngas (H₂/CO) ratio. The synergistic effect of endothermic and exothermic reactions in the system determines the overall energy input requirement and reaction kinetics. Low-energy electrons are mainly used for dissociation (En1–En4), medium-energy electrons (12–15 eV) promote ionization (En5–En6), and high-energy electrons (> 15 eV) trigger dissociation ionization (En7–En8). With the increase of electron temperature T_e, the ionization rate increases exponentially, while the rate of free radical recombination reaction decreases with the increase of gas temperature T_g.

Since DBD-DRM is a non-thermal reaction, the reaction rate does not fully conform to the Arrhenius formula, and it is necessary to introduce T_e to correct the rate. Fitting the electron impact rate constant k curve of CH₄ and CO₂ based on experimental data can be done by following the steps below:

Step 1: establish a kinetic model

Assuming that the main reaction path is electron impact dissociation (En2 and En4), the relationship between the reaction rate constant and the electron energy E_e is:

k=AE{}_{\mathrm{e}}.

Among them, A is the parameter to be fitted, and E_e is related to SEI.

Step 2: correlate SEI with electron energy E_e

According to the literature [10], there is a positive correlation between the electron temperature T_e (unit: eV) and SEI in DBD. Assumption:

E{}_{\mathrm{e}}=\alpha \times \mathrm{S}\mathrm{E}\mathrm{I},

where α is the proportional factor, which is calibrated by experimental data.

Step 3: fitting rate constant expression

I. Assuming first-order reaction kinetics:

X=1-e^{-kn{}_{\mathrm{e}}t}.

Where n_{\mathrm{e}} is the electron density and t is the residence time. Assuming n_{\mathrm{e}}t is a constant, then:

k\propto -\ln (1-X).

II. Calculate the k_{{\mathrm{C}\mathrm{H}}_4} ratio of CH₄ (verified by experimental data in Tab.1):

\frac{k_{240}}{k_{90}}=\frac{-\mathrm{l}\mathrm{n}(1-0.6126)}{-\mathrm{l}\mathrm{n}(1-0.3466)}=2.24.

Combined with k∝E_e and E_e∝SEI:

\frac{k_{240}}{k_{90}}=\sqrt{\frac{{SEI}_{240}}{{SEI}_{90}}}=1.63.

The experimental ratio (2.24) is higher than the theoretical value (1.63), indicating that the effect of SEI on the electron density ne needs to be introduced (such as n_{\mathrm{e}}∝SEI), then:

k\propto \sqrt{E{}_{\mathrm{e}}}\times n_{\mathrm{e}}\propto \sqrt{SEI}\times SEI=SE{I}^{1.5}.

The corrected ratio is:

\begin{array}{rl}& \frac{k_{240}}{k_{90}}={\left(\frac{240}{90}\right)}^{1.5}\approx 2.24\\ & \left(\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{i}\mathrm{s}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{w}\mathrm{i}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{x}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{u}\mathrm{e}\mathrm{s}\right).\end{array}

Therefore, the rate constant expression is:

k_{\mathrm{C}{\mathrm{H}}_4}=A\times SE{I}^{1.5}.

Step 4: determine parameter A

Using data when SEI = 90 J·mL^–1:

0.3466=1-e^{-kn{}_{\mathrm{e}}t}\Rightarrow k{n}_{\mathrm{e}}t\approx \mathrm{0.425.}

Assume n{}_{\mathrm{e}}t = B·SEI (because n{}_{\mathrm{e}}∝SEI), then:

k=A\times SE{I}^{1.5}\Rightarrow 0.425=A\times {90}^{1.5}\times B\times 90.

It needs to be solved in conjunction with CO₂ data, but for simplicity, A is calibrated directly:

A_{{\mathrm{C}\mathrm{H}}_4}=\frac{0.425}{{90}^{2.5}}\approx 6.2\times {10}^{-6}{\mathrm{c}\mathrm{m}}^3\cdot {\mathrm{s}}^{-1}\cdot {\mathrm{J}}^{-1.5}.

Similarly, a similar fitting is performed on the CO₂ conversion rate to obtain:

A_{{\mathrm{C}\mathrm{O}}_2}\approx 2.1\times {10}^{-6}{\mathrm{c}\mathrm{m}}^3\cdot {\mathrm{s}}^{-1}\cdot {\mathrm{J}}^{-1.5}.

Step 5: verify consistency with the literature

Combined with the electron collision cross section of CO₂ in the literature [15] ({\sigma }_{{\mathrm{C}\mathrm{O}}_2} = 3.2 × 10⁻¹⁶ cm²@ E_e = 6 eV):

k=\sigma \times v_{\mathrm{e}}\times n{}_{\mathrm{e}}=\sigma \times \sqrt{\frac{2eE{}_{\mathrm{e}}}{n{}_{\mathrm{e}}}}.

Substitute E_e = 6 eV, n_e = 10¹⁰ cm⁻³ (typical DBD value):

k_{{\mathrm{C}\mathrm{O}}_2}\approx 3.2\times {10}^{-16}\times 9.5\times {10}^7\times {10}^{10}=3.0\times {10}^{-3}{\mathrm{c}\mathrm{m}}^3\cdot {\mathrm{s}}^{-1}.

This is consistent with the fitting result (A_{{\mathrm{C}\mathrm{O}}_2}× 240^1.5 ≈ 2.1 × 10⁻⁶× 240^1.5 ≈ 0.008 cm³·s⁻¹).

k_{\mathrm{C}{\mathrm{H}}_4}=6.2\times {10}^{-6}\times SE{I}^{1.5},k_{\mathrm{C}{\mathrm{O}}_2}=2.1\times {10}^{-6}\cdot SE{I}^{1.5}.

The reaction rate constant (k\propto {SEI}^{1.5}) shows that the electron energy distribution and collision frequency affected by electron density (n_e) and T_e regulate the reaction efficiency through SEI. When SEI increases from 90 to 240 J·mL^–1, n_e and E_e increase simultaneously, significantly improving the CH₄ conversion rate (34.66% → 61.26%).

As SEI increases, T_e shows an upward trend, but the growth rate gradually slows down. At lower SEI (< 150 J·mL^–1), the electron temperature is low and the plasma reaction activity is limited (Fig.6). After SEI exceeds 180 J·mL^–1, T_e growth slows down, indicating that the electron energy may be close to the saturation value. The power increase may not significantly improve the methane conversion rate, but may promote side reactions (such as CO₂ excessive dissociation). The selection of a suitable SEI requires a balance between the electron temperature and the target reaction activity to avoid unnecessary energy loss and by-product generation. Therefore, precise control of the electron energy distribution and the gas phase temperature T_g is of great significance for optimizing the CH₄/CO₂ conversion reaction.

3.5 Ni-based catalysts in DBD-DRM

To investigate the performance and reaction mechanisms of Ni-based catalysts in DBD-DRM, 20 wt % Ni@γ-Al₂O₃ and 20 wt % Ni@SiO₂ are synthesized. The choice of 20 wt % Ni loading is based on a survey of previous studies, where Ni loadings in the range of 10–30 wt % have been commonly employed for plasma-assisted dry reforming of methane. Among them, 20 wt % Ni is widely reported to offer a good balance between metal dispersion, catalytic activity, and resistance to sintering under plasma conditions [35−37]. Moreover, the purpose of this study is not to develop novel catalysts, but to evaluate the kinetic behavior and plasma–catalyst interaction of representative and conventional Ni-based catalysts. Therefore, selecting a typical and well-studied loading such as 20 wt % enhances the comparability and reproducibility of the results, while providing a stable benchmark for investigating the effects of plasma parameters. The particle sizes of the catalysts vary, with γ-Al₂O₃ exhibiting a more uniform nanostructure (Fig.7(a) and 7(b)). The actual Ni loading on SiO₂ slightly exceeds the expected 20 wt %, as confirmed by EDS (Fig.7(c) and 7(d)). XRD shows that Ni is evenly distributed in both catalysts, without forming large crystalline particles and indicating a uniform atomic dispersion of Ni (Fig.7(e)). Raman scattering confirms no phase transformation of Ni, with characteristic peaks corresponding to Si–O and Al–O vibrations at 504 and 375 cm^–1, respectively (Fig.7(f)). This light primarily includes ultraviolet (UV), visible (Vis), and near-infrared (NIR) radiation, with wavelengths typically ranging from 100 to 3000 nm (Fig.7(g)). The specific wavelength range varies depending on factors such as the type of plasma, discharge power, and gas composition. UV light enhances reaction activity by exciting reactant molecules and influencing the optical properties of the catalyst. Visible light improves catalytic activity through surface plasmon resonance effects. NIR light manifests as thermal energy, suggesting that plasma discharge processes may facilitate light-assisted catalytic reactions. XPS reveals that Ni on SiO₂ has predominantly a higher oxidation state, indicating strong electronic interaction between Ni and SiO₂ and likely formation of NiO or Ni-Si-O compounds (Fig.7(h)–7(k)). In contrast, Ni on γ-Al₂O₃ exhibits a more complex chemical state, including both metallic Ni and Ni²⁺. The excitation and dissociation processes between electrons and gas molecules within the plasma generate electromagnetic radiation, resulting in a glow.

The H₂-TPR profiles of the fresh catalysts are shown in Fig.8(a). For the Ni@Al₂O₃ catalyst, two distinct hydrogen consumption peaks were observed at approximately 440 and 601 °C. The first peak can be attributed to the reduction of bulk NiO species that are weakly interacting with the Al₂O₃ support, while the second peak corresponds to the reduction of NiO species moderately interacting with the Al₂O₃ support. Notably, no high-temperature reduction peak was detected above 700 °C, suggesting that the formation of the strongly bound NiAl₂O₄ spinel phase is negligible under the applied preparation and calcination conditions. This indicates that NiO is the dominant Ni species present in Ni@Al₂O₃, and the metal-support interaction is moderate. In comparison, the H₂-TPR profile of the Ni@SiO₂ catalyst exhibited a single broad reduction peak centered at approximately 500 °C. This high-temperature peak can be ascribed to the reduction of NiO species confined within the porous SiO₂ framework or the formation of highly dispersed nickel silicate-like species, which are more difficult to reduce due to strong metal–support interactions and potential structural incorporation. The absence of a low-temperature peak suggests that weakly bound or bulk-like NiO is scarce in Ni@SiO₂, possibly due to strong dispersion or encapsulation of Ni species within the silica matrix. Clearly, the reduction temperature of Ni@Al₂O₃ is more distributed, with a noticeable fraction of NiO reducible at lower temperatures (440 °C), implying the existence of more easily reducible Ni species. In contrast, the single, higher-temperature peak in Ni@SiO₂ reflects a more uniform but strongly bound NiO phase. This indicates that Ni@Al₂O₃ has relatively better reducibility than Ni@SiO₂ under comparable conditions, which may lead to a higher fraction of active metallic Ni species being generated at moderate activation temperatures. Therefore, the support type plays a critical role in determining the reducibility of the Ni species. The Al₂O₃-supported Ni catalyst exhibited enhanced hydrogen activation and more favorable reduction characteristics compared to the SiO₂-supported counterpart. This difference may influence subsequent catalytic performance, especially in reactions where the generation of metallic Ni at lower temperatures is essential.

The CO₂-TPD profiles of Ni@Al₂O₃ and Ni@SiO₂ catalysts were analyzed to assess their CO₂ absorption capabilities, as illustrated in Fig.8(b). For Ni@SiO₂, the desorption peak observed at 202 °C corresponds to weak basic sites on the catalyst surface, while the peak at 428 °C is attributed to sites with moderate basicity. In contrast, Ni@Al₂O₃ exhibits a desorption peak at 601 °C, indicative of strong basic sites. The presence of strong basic sites in Ni@Al₂O₃ suggests a higher capacity for CO₂ chemisorption and dissociation compared to Ni@SiO₂, which primarily features weaker and moderately basic sites. This difference implies that Ni@Al₂O₃ would likely demonstrate superior performance in CO₂ conversion and methane activation due to its enhanced ability to adsorb and dissociate CO₂. Furthermore, the strong basic sites on Ni@Al₂O₃ can provide additional surface oxygen, improving the catalyst’s resistance to carbon deposition. In summary, the basicity and CO₂ absorption ability of the catalysts follow the sequence Ni@Al₂O₃ > Ni@SiO₂, with Ni@Al₂O₃ being more effective for reactions involving CO₂ due to its strong basic sites.

The catalytic performance of both catalysts increases with input power (Fig.9(a)). At 100 W, 20 wt % Ni@SiO₂ exhibits a slightly higher CO₂ conversion rate (25.36%) compared to 20 wt % Ni@γ-Al₂O₃ (21.78%), possibly due to enhanced exposure of active sites on SiO₂ at high power. At 40 W, the difference in conversion rates between the two catalysts is minimal, indicating limited activation of active sites. Similarly, the CH₄ conversion rate increases with power, with 20 wt % Ni@SiO₂ showing slightly higher conversion at 100 W (62.55%) compared to 20 wt % Ni@γ-Al₂O₃ (62.05%), likely due to the better dispersion of Ni particles and more uniform distribution of active sites on SiO₂.

The production rates of H₂ and CO exhibit different trends. At 40 W, the H₂ yield from 20 wt % Ni@γ-Al₂O₃ is slightly higher, but as power increases, the yields of both catalysts approach each other (Fig.9(b)). Notably, the CO yield of 20 wt % Ni@SiO₂ is higher than that of 20 wt % Ni@γ-Al₂O₃ under all power conditions, especially 100 W because the neutral surface characteristics of SiO₂ suppresses the water-gas shift reaction, leading to higher CO retention. 20 wt % Ni@SiO₂ exhibits slightly better selectivity for H₂ and CO at high power, while 20 wt % Ni@γ-Al₂O₃ shows higher selectivity for hydrocarbon byproducts (such as C₂H₄, C₂H₆, and C₃H₈) (Figs. S4(a, b), cf. ESM). This may be because the basic surface sites on γ-Al₂O₃ promote hydrocarbon polymerization. Both catalysts maintain a high carbon balance close to 99%–100% (Fig.9(c)), indicating minimal carbon loss and high system efficiency during DBD-DRM. The carbon balance of 20 wt % Ni@SiO₂ is slightly higher than that of 20 wt % Ni@γ-Al₂O₃, suggesting better suppression of side reactions, particularly coke formation. However, a closer examination of the data reveals subtle but consistent differences. For instance, at a CH₄/CO₂ ratio of 40, the carbon balance of 20 wt % Ni@γ-Al₂O₃ is 98.36%, noticeably lower than the 99.93% observed for 20 wt % Ni@SiO₂. Although the carbon balance of Ni@γ-Al₂O₃ improves at higher ratios (e.g., up to 99.74% at 100), the values remain slightly lower overall compared to Ni@SiO₂ across all test points. These results suggest that Ni@SiO₂ exhibits better suppression of carbon loss pathways, likely due to its weaker surface basicity, which reduces hydrocarbon polymerization and subsequent coke formation. In contrast, the stronger basicity of γ-Al₂O₃ may favor C–C coupling reactions and polymer growth, leading to minor but measurable carbon retention on the catalyst or reactor surfaces. Therefore, while both catalysts perform well in maintaining a high carbon balance, Ni@SiO₂ shows slightly superior control over side reactions, which may be advantageous for long-term plasma-catalytic DRM applications.

Under reaction conditions of CO₂:CH₄:He = 6:6:48, both the CO₂ and CH₄ conversion rates increase with power. At low power, the catalytic effect of the active metal is more pronounced, while at higher power, the physical properties of the substrate have a greater influence on the reaction efficiency (Fig.9(d)). As the power increases from 40 to 100 W, the conversion rates of all the catalyst-substrate combinations improve due to enhanced plasma activation, which produces better excitation and decomposition of reactant molecules.

The differences between catalysts and substrates are significant. When using quartz sand alone, the conversion rates at high power are similar to those of the mixed systems. Particularly at 80 and 100 W, quartz sand delivers excellent performance, likely due to its high-temperature stability and large surface area. When mixed with 20 wt % Ni@γ-Al₂O₃, quartz sand dominates the reaction at higher power, diluting the activity of the Ni-based catalyst [11]. In contrast, the combination of 20 wt % Ni@SiO₂ and quartz sand shows optimal results at 60 and 80 W, suggesting that the SiO₂ substrate facilitates the uniform distribution of metal particles and efficient utilization of active sites. Using γ-Al₂O₃ alone results in the lowest conversion rate at low power, but at high power, its performance approaches that of other mixed systems, indicating that plasma activation can compensate at lower power. The results show that different catalyst-substrate combinations are suitable for different power conditions, with the dispersion of the catalyst and the material properties of the substrate playing key roles in the conversion rates, especially at low input power.

This study shows that different catalyst and substrate combinations are suitable for different power conditions. Under high power conditions, since the catalytic reaction of DBD is close to the thermodynamic limit, Ni-based catalysts have little effect on improving the conversion rate. On the contrary, an efficient substrate, such as quartz sand, can improve the discharge effect. Under low power conditions, due to the limited input energy, the dispersion of metal catalysts and the selection of substrate both affect the improvement of reaction efficiency. Combined with the analysis of energy conversion rate and E_eff, in order to reduce the energy consumption of DBD reactors, future research on high CO₂ and CH₄ conversion rates under low input power mode should focus on the dispersion of catalysts and the material properties of carriers.

To clarify, the primary objective of this study is to investigate the catalytic performance and reaction behavior of conventional Ni-based catalysts under DBD-assisted dry reforming conditions, with a focus on the effects of discharge parameters, gas composition, and catalyst type on conversion efficiency and product distribution. Although catalyst stability is an important parameter for practical application, the current reactor configuration imposes certain limitations for long-duration testing. Specifically, under extended operation (beyond 10 h), localized hotspots caused by plasma–catalyst interactions may lead to gradual temperature rise, catalyst fouling, and the formation of condensed organic residues on the reactor wall or electrode surface, which ultimately affect discharge uniformity and reaction reliability.

Given these constraints, this work does not include long-term durability studies. Instead, it emphasizes short-term performance trends and mechanistic insights under varying process conditions. In future research, we plan to optimize the DBD reactor design and its thermal management system to enable extended, stable plasma-catalytic operation and allow for comprehensive lifetime evaluation of catalyst performance.

4 Conclusions

This work systematically evaluated the catalytic performance of the DBD-DRM system and deeply explored the effects of reaction parameters such as electric field intensity, gas flow rate, and temperature on the DRM reaction efficiency. Under the action of DBD, Ni-based catalysts showed excellent catalytic activity and carbon deposition resistance, and the catalytic performance was closely related to Ni particle size, dispersion, and metal-support interaction. The 20 wt % Ni@SiO₂ catalyst showed a higher conversion rate and product selectivity due to its larger pore size, higher pore volume, and excellent Ni particle dispersion. 20 wt % Ni@γ-Al₂O₃ showed better catalytic stability due to its strong metal-support interaction. The high-energy electrons and active free radicals generated by DBD effectively promoted the decomposition of CH₄ and CO₂ through vibrational excitation and direct collision. By constructing a nonequilibrium thermodynamic model, the energy distribution and competition mechanism of CH₄/CO₂ dry reforming in the DBD system were revealed. The high consistency between experimental data and theoretical predictions verified the reliability of the model. This work not only deepens the understanding of non-thermal plasma chemistry, but also provides technical support for the resource utilization of methane in the context of carbon neutrality.

References

[1] Wang H , Diao Y , Gao Z , Smith K J , Guo X , Ma D , Shi C . H₂ production from methane reforming over molybdenum carbide catalysts: from surface properties and reaction mechanism to catalyst development. ACS Catalysis, 2022, 12(24): 15501–15528

[2] Mohanty U S , Ali M , Azhar M R , Al-Yaseri A , Keshavarz A , Iglauer S . Current advances in syngas (CO + H₂) production through bi-reforming of methane using various catalysts: a review. International Journal of Hydrogen Energy, 2021, 46(65): 32809–32845

[3] Yentekakis I V , Panagiotopoulou P , Artemakis G . A review of recent efforts to promote dry reforming of methane (DRM) to syngas production via bimetallic catalyst formulations. Applied Catalysis B: Environmental, 2021, 296: 120210

[4] Yu J , Le T , Jing D , Stavitski E , Hunter N , Lalit K , Leshchev D , Resasco D E , Sargent E H , Wang B . . Balancing elementary steps enables coke-free dry reforming of methane. Nature Communications, 2023, 14(1): 7514

[5] Zhu Q , Zhou H , Wang L , Wang L , Wang C , Wang H , Fang W , He M , Wu Q , Xiao F S . Enhanced CO₂ utilization in dry reforming of methane achieved through nickel-mediated hydrogen spillover in zeolite crystals. Nature Catalysis, 2022, 5(11): 1030–1037

[6] Chan Y H , Chan Z P , Lock S S M , Yiin C L , Foong S Y , Wong M K , Ishak M A , Quek V C , Ge S , Lam S S . Thermal pyrolysis conversion of methane to hydrogen (H₂): a review on process parameters, reaction kinetics, and techno-economic analysis. Chinese Chemical Letters, 2024, 35(8): 109329

[7] Yu K , Wang C , Zheng W , Vlachos D G . Dynamic electrification of dry reforming of methane with in situ catalyst regeneration. ACS Energy Letters, 2023, 8(2): 1050–1057

[8] Nozaki T , Chen X , Kim D Y , Zhan C . Combination of DBD and catalysts for CH₄ and CO₂ conversion: basics and applications. Plasma Chemistry and Plasma Processing, 2023, 43(6): 1385–1410

[9] García-Moncada N , Van Rooij G , Cents T , Lefferts L . Catalyst-assisted DBD plasma for coupling of methane: minimizing carbon-deposits by structured reactors. Catalysis Today, 2021, 369: 210–220

[10] Aerts R , Martens T , Bogaerts A . Influence of vibrational states on CO₂ splitting by dielectric barrier discharges. Journal of Physical Chemistry C, 2012, 116(44): 23257–23273

[11] Ray D , Nepak D , Janampelli S , Goshal P , Subrahmanyam C . Dry reforming of Methane in DBD plasma over Ni-based catalysts: influence of process conditions and support on performance and durability. Energy Technology, 2019, 7(4): 1801008

[12] Wang H , Yang Y , Li Z , Kong X , Martin P , Cui G , Wang R . Plasma-assisted Ni catalysts: toward highly-efficient dry reforming of methane at low temperature. International Journal of Hydrogen Energy, 2023, 48(24): 8921–8931

[13] Tu X , Whitehead J . Plasma-catalytic dry reforming of methane in an atmospheric dielectric barrier discharge: understanding the synergistic effect at low temperature. Applied Catalysis B: Environmental, 2012, 125: 439–448

[14] Wang K , Ren X , Yin G , Hu E , Zhang H . Recent advances in plasma-based methane reforming for syngas production. Current Opinion in Green and Sustainable Chemistry, 2024, 50: 100981

[15] Bogaerts A , Kozák T , Van Laer K , Snoeckx R . Plasma-based conversion of CO₂: current status and future challenges. Faraday Discussions, 2015, 183: 217–232

[16] Snoeckx R , Bogaerts A . Plasma technology—a novel solution for CO₂ conversion. Chemical Society Reviews, 2017, 46(19): 5805–5863

[17] Snoeckx R , Zeng Y , Tu X , Bogaerts A . Plasma-based dry reforming: improving the conversion and energy efficiency in a dielectric barrier discharge. RSC Advances, 2015, 5(38): 29799–29808

[18] Khoja A H , Mazhar A , Saleem F , Mehran M T , Naqvi S R , Anwar M , Shakir S , Saidina Amin N A , Sajid M B . Recent developments in catalyst synthesis using DBD plasma for reforming applications. International Journal of Hydrogen Energy, 2021, 46(29): 15367–15388

[19] Li J Q , Bin X , Wang W B , Xie J J , Yin X L , Wu C Z , Xiao J B . Experimental study on dry reforming of methane by a plasma catalytic hybrid system. Journal of Fuel Chemistry & Technology, 2021, 49(8): 1161–1172

[20] Uytdenhouwen Y , Hereijgers J , Breugelmans T , Cool P , Bogaerts A . How gas flow design can influence the performance of a DBD plasma reactor for dry reforming of methane. Chemical Engineering Journal, 2021, 405: 126618

[21] Lu Q , Lei W , Yue W , Huang W , Dong Y , Yan W , Liu Y , Chen Y , Zhao Y . Reaction mechanism explorations on non-thermal plasma reforming of CO₂-CH₄ by combining kinetics modeling and emission spectroscopy measurements. Fuel, 2023, 344: 128041

[22] Baig S , Sajjadi B . Non-thermal plasma enhanced catalytic conversion of methane into value added chemicals and fuels. Journal of Energy Chemistry, 2024, 97: 265–301

[23] Sun J , Chen Q , Qin W , Wu H , Liu B , Li S , Bogaerts A . Plasma-catalytic dry reforming of CH₄: effects of plasma-generated species on the surface chemistry. Chemical Engineering Journal, 2024, 498: 155847

[24] Arshad M Y , Saeed M A , Tahir M W , Pawlak-Kruczek H , Ahmad A S , Niedzwiecki L . Advancing sustainable decomposition of biomass tar model compound: machine learning, kinetic modeling, and experimental investigation in a non-thermal plasma dielectric barrier discharge reactor. Energies, 2023, 16(15): 5835

[25] Sanjaya E , Weihs G F , Manaf N A , Abbas A . Reaction kinetic modelling of tar cracking in a non-thermal plasma reactor: model evaluation and reaction mechanism investigation. Chemical Engineering Journal, 2024, 479: 147649

[26] Wang J , Zhang K , Bogaerts A , Meynen V . 3D porous catalysts for plasma-catalytic dry reforming of methane: how does the pore size affect the plasma-catalytic performance. Chemical Engineering Journal, 2023, 464: 142574

[27] Miralai S , Monette E , Bartnikas R , Czeremuszkin G , Latreche M , Wertheimer M R . Electrical and optical diagnostics of dielectric barrier discharges (DBD) in He and N₂ for polymer treatment. Plasmas and Polymers, 2000, 5(2): 63–77

[28] Chiper A S , Aniţa V , Agheorghiesei C , Pohoaţa V , Aniţa M , Popa G . Spectroscopic diagnostics for a DBD plasma in He/Air and He/N₂ gas mixtures. Plasma Processes and Polymers, 2004, 1(1): 57–62

[29] Andersen J , Christensen J , Østberg M , Bogaerts A , Jensen A . Plasma-catalytic dry reforming of methane: screening of catalytic materials in a coaxial packed-bed DBD reactor. Chemical Engineering Journal, 2020, 397: 125519

[30] Van Turnhout J , Aceto D , Travert A , Bazin P , Thibault-Starzyk F , Bogaerts A , Azzolina-Jury F . Observation of surface species in plasma-catalytic dry reforming of methane in a novel atmospheric pressure dielectric barrier discharge in situ IR cell. Catalysis Science & Technology, 2022, 12(22): 6676–6686

[31] Wang B , Li X , Wang X , Zhang B . Effect of filling materials on CO₂ conversion with a dielectric barrier discharge reactor. Journal of Environmental Chemical Engineering, 2021, 9(6): 106370

[32] Lü F C , Ruan H O , Song J X , Yin K , Zhan Z Y , Jiao Y F , Xie Q . Enhanced surface insulation and depressed dielectric constant for Al₂O₃/epoxy composites through plasma fluorination of filler. Journal of Physics D: Applied Physics, 2019, 52(15): 155201

[33] Jiang N , Qu Y , Zhu J , Wang H , Li J , Shu Y , Cui Y , Tan Y , Peng B , Li J . Remediation of atrazine-contaminated soil in a fluidized-bed DBD plasma reactor. Chemical Engineering Journal, 2023, 464: 142467

[34] Chi J , Wu X , Tao J , Min X , Li Z , Zhao S . DBD-coupled highly dispersed Ni/SiO₂ materials for CO₂ reduction performance and mechanism study. Journal of the Energy Institute, 2024, 113: 101552

[35] Mei D , Tu X . Atmospheric pressure non-thermal plasma activation of CO₂ in a packed-bed dielectric barrier discharge reactor. ChemPhysChem, 2017, 18(22): 3253–3259

[36] Zhang K , Zhang G , Liu X , Phan A N , Luo K . A study on CO₂ decomposition to CO and O₂ by the combination of catalysis and dielectric-barrier discharges at low temperatures and ambient pressure. Industrial & Engineering Chemistry Research, 2017, 56(12): 3204–3216

[37] Sivachandiran L , Da Costa P , Khacef A . CO₂ reforming in CH₄ over Ni/γ-Al₂O₃ nano catalyst: effect of cold plasma surface discharge. Applied Surface Science, 2020, 501: 144175

RIGHTS & PERMISSIONS

The Author(s) 2025. This article is published with open access at link.springer.com and journal.hep.com.cn